Machine learning allows ultrafast AFM atomic measurements

October 09, 2017 // By Nick Flaherty
Researchers at the Oak Ridge National Laboratory in the US have developed a new technique for high speed voltage measurements at the atomic level using machine learning.

The researchers developed a technique for making ultrafast measurements using atomic force microscopy, which previously could only investigate slow or static material structures and functions. In AFM, a rastering probe maps a material's surface and captures physical and chemical properties, but the probe is slow to respond to what it detects.

Instead, the ORNL fast free force recovery technique uses advanced machine learning algorithms to analyse instantaneous tip motion to produce high-resolution images 3,500 times faster than standard AFM detection methods.

"This new approach can probe fast processes, such as charge screening, ionic transport and electrochemical phenomena, which were previously inaccessible with traditional AFM," said Liam Collins at ORNL. 

The Fast free force recovery (F 3R) method allow electrostatic forces to be measured with a temporal resolution of 10μs. When this is applied to open loop Kelvin probe force microscopy (KPFM) measurements without bias feedback on the AFM tip, this allows ultrafast surface potential measurements under 20 μs to be performed at regular KPFM scan speeds. The technique has been used to map the surface voltage dynamics from ion migration induced by an electric field in a perovskite solar cell (shown above).

Related stories:


Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.